Scholarlink

<p>Despite the increase in power generation through centralized and distributed energy sources, the imbalances between energy supply and demand still exist. This study utilizes permissioned blockchain for peer-to-peer energy trading and a predictive analytics approach for making accurate predictions on each user’s energy consumption and energy generation for minimizing this disparity with optimized energy transfer. The four models include Long Short-Term Memory, Bidirectional Long Short-Term Memory, Gated Recurrent Unit, and Prophet by Facebook. These models are tested on Germany’s household and industrial energy datasets. A comparison among the four predictive models is done on the mean absolute error and root mean square error performance metrics. The performance evaluation of the blockchain system is performed using the Hyperledger Caliper tool with the number of transactions ranging between 50 and 1500. The maximum throughput and average latency achieved for the transaction function is 32.2 transactions per second and 20 milliseconds, whereas for the query function, it is 27.5 transactions per second and 23.3 milliseconds. A comparative analysis with the previous related research is also provided. These experimentation results represent the effectiveness and practicality of the proposed framework.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

FEED: Fair and Equitable Energy Distribution based on permissioned blockchain and predictive analytics

  • 期刊论文

  • 发表日期：

    2025年7月31日

查看全文
  • Rohit Shorya,
  • Priti Jagwani

摘要

Despite the increase in power generation through centralized and distributed energy sources, the imbalances between energy supply and demand still exist. This study utilizes permissioned blockchain for peer-to-peer energy trading and a predictive analytics approach for making accurate predictions on each user’s energy consumption and energy generation for minimizing this disparity with optimized energy transfer. The four models include Long Short-Term Memory, Bidirectional Long Short-Term Memory, Gated Recurrent Unit, and Prophet by Facebook. These models are tested on Germany’s household and industrial energy datasets. A comparison among the four predictive models is done on the mean absolute error and root mean square error performance metrics. The performance evaluation of the blockchain system is performed using the Hyperledger Caliper tool with the number of transactions ranging between 50 and 1500. The maximum throughput and average latency achieved for the transaction function is 32.2 transactions per second and 20 milliseconds, whereas for the query function, it is 27.5 transactions per second and 23.3 milliseconds. A comparative analysis with the previous related research is also provided. These experimentation results represent the effectiveness and practicality of the proposed framework.